Sweet potato (Ipomoea batatas L.) is a nutritious food crop and, based on the high starch content of its storage root, a potential bioethanol feedstock. Enhancing the nutritional value and starch quantity of storage roots are important goals of sweet potato breeding programs aimed at developing improved varieties for direct consumption, processing, and industrial uses. However, developing improved lines of sweet potato is challenging due to the genetic complexity of this plant and the lack of genome information. Short sequence repeat (SSR) markers are powerful molecular tools for tracking important loci in crops and for molecular-based breeding strategies; however, few SSR markers and marker-trait associations have hitherto been identified in sweet potato. In this study, we identified 1,824 SSRs by using a de novo assembly of publicly available ESTs and mRNAs in sweet potato, and designed 1,476 primer pairs based on SSR-containing sequences. We mapped 214 pairs of primers in a natural population comprised of 239 germplasms, and identified 1,278 alleles with an average of 5.972 alleles per locus and a major allele frequency of 0.7702. Population structure analysis revealed two subpopulations in this panel of germplasms, and phenotypic characterization demonstrated that this panel is suitable for association mapping of starch-related traits. We identified 32, 16, and 17 SSR markers associated with starch content, β-carotene content, and starch composition in the storage root, respectively, using association analysis and further evaluation of a subset of sweet potato genotypes with various characteristics. The SSR markers identified here can be used to select varieties with desired traits and to investigate the genetic mechanism underlying starch and carotenoid formation in the starchy roots of sweet potato.